2. Big buttons work. They are the call to action, and big call to actions get a big amount of attention.
3. When picking images, be mindful of the market you are selling to. The race of the people in the images are very important. Not because of racism, but due to how people relate to those who look like themselves. A self-taken photograph usually nets better results than stock photos. Even one from a mobile phone.
4. If you are selling a service, then include a photo of the dashboard/main area in the main page. People want to see how it looks like without signing up.
All one needs to do is consult the artist within and choose the right image for better conversion. That's exactly the point no. 3, like you said, kudos. IMHO, images and their messaging are probably one of the spots where art and math can converge.
I may be combining your first and second sentences together in a way that you did not intend, but I think the adage is older than scrolling ...
In general, though, I do completely agree with the article that the only way to find out what works is to test, test, test.
It's the cargo cult approach to testing: "let's change everything and if the revamp increases conversions we'll guess what factor had the most effect".
Also I think it is important to include feedback from actual users in your test results. How did they feel using the different sites? What are the users feelings about your own concerns (invisible content below the fold, ridiculous big buttons, etc)?
Without these considerations you are just poking around the mud with a stick.
Is this really an "understood" truth? In our own study for our own site, we found nearly everyone scrolled the page all the time, and that seems the be the case with every colleague I've ever talked to as well. Who goes to a page and says, "fuck it, that scroll wheel is just way too damned difficult"?
The thing that is puzzling to me is the whole thing about people scrolling & responsive design. How do you design something to avoid scrolling when your layout is designed to accommodate any screen size? It seems like those concepts are in conflict.
However, as the author admits in the comments (after being corrected by a reader), he has interpreted the results of the source incorrectly. The cited report actually only says that "Web users spend 80% of their time looking at information above the page fold. Although users do scroll, they allocate only 20% of their attention below the fold."
My guess is: if you capture their curiosity, they will scroll. Otherwise, they probably won't. Simple, I know, but I'm going with it.
In the other cases, it seems more like "a change happened" and "conversion rates went up" - far too many changes at once to imply one specific change caused it :-/
It's a big differentiator when the same (or equivalent) product is sold by multiple merchants. Customers in my experience are even willing to pay a premium to the merchant with the better content, presumably because they seem more legitimate and less-risky.
So try it for yourself. Track results. And make sure you do the statistics right to know if you're looking at signal or noise. (Early noise can be very, very impressive.)
At first blush all but the last Dell treatment are absolutely horrendous to my eye, so I just don't know how to even judge these.
Everybody should take a college level course in statistics, even poets and painters.
Looking at how some really great sites are doing, too, there are sites with massive pictures, but if they're not high quality, turns out to look like a cheap site.
There's only one sentence on that page and, unless there's a new term with which I'm not familiar, it has one heck of a credibility-destroying typo:
"Get our converison [sic] tips newsletter starting now:"